International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
Volume: 4 Issue: 4
162 - 163
___________________________________________________________________________________________________________________

Review on Automatic and Customized Itinerary Planning Using Clustering

Algorithm and Package Recommendation for Tourism Services
Sneha Shahane1, B. R. Solunke2
1
2

M.E.(CSE), Department of Computer Science and Technology, NBNSCOE, Solapur University, Solapur, India
Asst. Prof., Department of Computer Science and Technology, NBNSCOE, Solapur University, Solapur, India

Abstract Itinerary planning is a strongly studied problem concerned with how to create the most optimum itinerary for a user taking into
account their various preferences. There are number of constraints which are to be considered such as maximum monetary value user willing for
paying, maximum duration to spend, and details about different POIs. Since the search space of all probable itineraries is too costly to fully
explore to make straightforward the complexity most work assume that user's trip is incomplete to some significant POIs and will complete
within a days. Therefore to overcome this problem, in this service, each POI is considered and ranked based on the users preferences. Since the
many users are planning for a trip with different requirements then their search complexity is increased. To overcome this limitation, the concept
of grouping or clustering the users based on users requirement similarity is proposed. To reduce the processing cost, a two-stage planning
scheme is been proposed. These are preprocessing stage and online stage. This paper also proposes the work that concentrates on recommending
an itinerary to a user given the users travel preferences and constraints. Finally packages containing optimum tour path satisfying customer
needs are generated efficiently.
Keywords team orienteering problem, itinerary planning, clustering, recommendation

__________________________________________________*****_________________________________________________
1. INTRODUCTION
Planning a trip can take a lot of time and effort. One needs to
explore points of interest (POIs) that one wants to visit, where
they want to stay, and how they want to travel. They may also
have restrictions on how much time they can spend at a
locations, and what their budget for the trip is. Among the
available POIs, there can be certain categories of attractions
that a user would not like, or perhaps they have nonmainstream interests. Even though the travel agency makes
available efficient and expedient services for experienced
travelers the itineraries provided by the travel agency lack
customization and cannot fulfill individual provisions.
Therefore, to attract more customers, travel agency
should permit the users to customize their itineraries. The
itinerary planning service returns top-K trip plans satisfying
the requirements. In this service, all POIs are considered and
ranked based on the users preference.
Present planning algorithms only consider a single
days trip, whereas in real cases; most users will schedule an
n-day itinerary. The travel agents tend to favor the popular
POIs. Suppose we have N available POIs and there are m POIs
in each single days itinerary averagely. It becomes expensive
to evaluate benefit of each itinerary and choose optimal one.
Also handling new emerging POIs was tricky in previous
approaches.
To overcome such problems, here[1], a novel
itinerary planning approach is proposed. The design
philosophy of proposed approach is to generate itineraries that
reduce the gap between the agents and travelers. Hence this
approach can be summarized as follows.
Searching the optimal itinerary can be converted into
the team orienteering problem (TOP), which is an NPcomplete problem without polynomial approximations. Hence,
a two stage scheme is applied. In the preprocessing stage,
POIs are organized into an undirected graph, G also all
candidate single-day itineraries are iterated using a parallel
processing framework, MapReduce. In the online stage, design

an approximate algorithm is designed to generate the optimal

itineraries. The approximate algorithm adopts the
initialization-adjustment model.
Also here another method for efficient itinerary planning
clustering algorithm such as FKM Clustering or nearest
neighbor clustering is used. Clusters of Relevant users and
POIs are made.
This work also aims to explore using top-k package
selection to select a set of POIs that will adhere to the users
preferences, time, and money constraints. There are three
different stages to the itinerary recommendation process. The
first stage is to train the algorithm by finding user preferences
from previous trips that the user took. The second stage is the
top k package recommendation where sets of POIs are selected
to create a package. Finally, the third stage is to order the POIs
logically, to create an itinerary.
At last, various packages are generated and
recommended for different users depending on their various
requirements.
2. RELATED WORK
Following are some pre-existing studies and approaches
related for automatic itinerary planning, clustering, and
package recommendation:
2.1 Grouping and Clustering
In [2], R.Rajeswari proposed the concept of grouping or
clustering the users dependent on users requirement
similarity. The novel modified clustering is been used for
grouping the user as per their requirement similarity. Here, the
POIs that user wants to visit are gathered, after which the
location information from the users, K-means clustering has
been applied to group the nearest locations together. The
membership values are cleared for clustering process and its
based on the users. It is used to reduce the search complexity
as well as time complexity of the system.
162

IJRITCC | April 2016, Available @ http://www.ijritcc.org

________________________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
Volume: 4 Issue: 4
162 - 163
___________________________________________________________________________________________________________________

2.2 Recommendation of Top-k Packages

SampoornaBiwas in [3] aims on recommending an itinerary to
a user, depending on given users travel preferences and
constraints. This scheme mainly aims to explore using top-k
package selection to select a set of POIs that will adhere to the
users preferences, time, and money constraints. The itinerary
algorithm used here expects inputs as-maximum monetary
value the user is willing to spend on tour, Maximum duration,
details about various types of attractions and place user wants
to visit. Three stages of itinerary recommendation process are
also mentioned. Initially algorithm is trained by finding user
preferences from history of user. Next top-k package
recommendation is done where sets of POIs are selected to
generate a package. Finally, in last stage POIs are ordered
logically, to form an itinerary.
2.3 Recommendation Using Flickr
In [4], Steven Van Canneyt, propose a system that
recommends tourist attractions based on the moment that the
user visits a given city. They started from a large collection of
georeferenced photos on Flickr, and use Mean Shift clustering
to determine points of interest within a city. Then they
estimated the probability that a random user would visit a
given place within a given temporal context.
Here, the possibility of using temporal context factors to
better predict is been explored which POIs might be of
interest to a given user.The experimental results determine
that significantimprovements over this baseline system
whose only criterion is the complete popularity of the place
can be obtained.
2.4 Greedy Local Improvement and Weighted Set Packing
Approximation [5]
Here presented an approximation algorithm for the weighted
k-set packing problem that combines two paradigms by
starting with a first greedy solution and then repeatedly
choosing the best possible local improvement. Considering
neighborhoods of sets of vertices in local improvement stage
reduce the additional term in performance ratio.
2.5 MapReduce: Data Processing Tool
Dean J. [6] has described the details of Mapreduce. Mapreduce
is a programming model for processing and generating large
data sets.Users define a map function that processes a key
/value pair to generate a set of intermediate key /value pairs
and a reduce function that integrates all intermediate values
associated with the same intermediate key. Mapreduce gives
many significant advantages over parallel databases.
2.6 Metaheuristics for the Team Orienteering Problem[7]
The extension of the Orienteering Problem to the case
of multiple tours is known as the Team Orienteering Problem
(TOP).Here among the metaheuristics proposed for the
solution of combinatorial optimization problems, tabu search
has been shown to be best for vehicle routing problems.

Another goodmetaheuristic is the variable neighborhood

search.
3. CONCLUSION
When planning a trip, tourists often want to avoid the
headache and time cost of having to slow research and plan
their trip. However, prepackaged itineraries often dont take
their own travel preferences, time cost, and monetary budget
into account. The service creates a customized multiday
itinerary dependent on the multiple users preference. As there
need to search for the optimal solution, a two-stage scheme is
used. In the preprocessing stage, iterate and index the
candidate single-day itineraries by using mapreduce
framework. After the preprocessing stage, the TOP is
transformed into weighted set-packing problem, which has
efficient approximate algorithms. In next stage, simulate the
approximate algorithm for set-packing problem. The algorithm
follows the initialization- adjustment model and can generate a
result. By using the top-k package generation algorithm and
package ordering, algorithms considering time and money
constraints as well as travel preferences in generated
itineraries.
4.

REFERENCES

[1] G. Chen, S. Wu, J. Zhou, and A. K. Tung.,Automatic

itinerary planning for traveling
services, Knowledge
and Data Engineering, IEEE Transactions on, 2014.
[2] Efficient Multiuser Itinerary Planning For Travelling
Services Using Fkm-Clustering Algorithm,R.Rajeswari, J.
MannarMannan, IRJET, May 2015
[3] SampoornaBiwas, Sam Creed, Jessica Wong, Itinerary
Planning Using Top-k Package Recommendation and
Multiple Constraints
[4] Steven Van Canneyt, Steven Schockaert , Olivier Van
Laere ,Bart Dhoedt, Time-Dependent Recommendation of
Tourist Attractions using Flickr
[5] Halldorsson M.M. and Chandra B., (2001), Greedy Local
Improvement and Weighted Set Packing Approximation, J.
Algorithms, Vol. 39, pp. 223-240.
[6] Dean J. and Ghemawat S., (2010), MapReduce: A Flexible
Data Processing Tool, Comm. ACM, Vol. 53, pp. 72-77.
[7] Archetti C., Hertz A. and Speranza M.G., (2007),
Metaheuristics for the Team Orienteering Problem, J.
Heuristics, Vol. 13, pp. 49-76.

AUTHORS
Sneha S. Shahane received the B.E. degree in Computer
Science and Engineering from Walchand institute of
Technology in 2014 and is pursuing M.E. in Computer
Science and Engineering from N. B. N. Sinhgad College of
Engineering, Solapur, India.
Prof.
B.
R.
SolunkecompletedM.Tech.(CSE)Shivaji
University, Kolhapur and working as Assistant Professor,
Department of Computer Science and Engineering, N. B. N.
Sinhgad College of Engineering, Solapur, India.

163
IJRITCC | April 2016, Available @ http://www.ijritcc.org

________________________________________________________________________________________________________